Background and Study Aims: Hemorrhage is induced in approximately 0.5-12% o
f endoscopic biliary sphincterotomy (ES) procedures. We prospectively inves
tigated the risk factors for ES-induced hemorrhage and evaluated the safety
as well as the effectiveness of endoscopic hemostasis.
Patients and Methods: The study included 1304 patients who underwent ES bet
ween July 1996 and June 1998. Epinephrine spray was used initially for hemo
static treatment. If bleeding persisted, an epinephrine injection was given
consecutively. In patients with exposed vessels, epinephrine injection fol
lowed by alcohol injection was given.
Results: ES-induced hemorrhage occurred in 136 (10.4 %) patients. The type
of sphincterotome used (needle-knife sphincterotome, P = 0.025) and the cut
ting speed (the so-called "zipper" cut, P = 0.049) were revealed as signifi
cant variables with regard to the occurrence of bleeding. Mild, moderate, a
nd severe bleeding were noted in 108 (79.4 %), 22 (16.2 %), and six (4.4 %)
patients, respectively. Once bleeding occurred, patients with an associate
d ampullary lesion (impacted stone or cancer) or with coagulopathy were mor
e likely to bleed profusely. Initial hemostasis was achieved in all patient
s. However, rebleeding occurred in eight patients who were initially classi
fied as having moderate or severe bleeding. Finally, ES-induced hemorrhage
was successfully controlled in all patients after 1-3 treatment sessions (m
ean 1.1 sessions). The difference in the incidence of complications between
the groups treated or not treated by endoscopic hemostasis was not statist
ically significant.
Conclusions: ES-induced hemorrhage occurred in 10% of the patients studied.
The use of needle-knife sphincterotomy and the cutting speed were independ
ent risk factors for the occurrence of bleeding. Once bleeding occurred, it
s severity was affected by an associated ampullary lesion (impacted stone o
r cancer) or coagulopathy. Endoscopic hemostasis with epinephrine and/or al
cohol was effective and safe in ES-induced hemorrhage.